Re: PATCH to document init_xxx and other function naming conventions

[Gerald Pfeifer <gerald at pfeifer dot com>]

On Sat, 9 Aug 2003, Joseph S. Myers wrote:
> Without a consensus on what goes where I don't want a patch that that
> spreads the information about more places than at present or that would
> be part of an incomplete transition without consensus to complete it.
> [...]
> Putting this documentation in codingconventions.html pending such an
> analysis and discussion would be fine.)

On Sat, 9 Aug 2003, Per Bothner wrote:
> I understand that.  On the other hand, it seems clear to me that
> documenting function naming standards is helpful to understanding
> the compiler, both for compiler "readers" and "writers", and so
> clearly belongs in the internals manual.  Exactly where in the
> manual to put it I'm less sure about.

I'm afraid we are stalled here?

What I'm seeing is a volunteer (Per) that wrote down a nice and useful
description of a previously undocumented part of our coding standards,
and dissens on where to put that.

Dissens on where to put that documentation, however, really shouldn't
prevent the documentation to be accepted in the long term (though I
understand part of Joseph's worries).

Would comitting the part of the patch that refers to the web page (just
codingconventions.html to keep the number of possibly broken links small)
and the part that specifies the function names and adding the following
part to

 +Code should follow ISO C89--GCC can no longer be compiled
 +using old pre-ANSI ``K&R'' C compilers.
 +All functions must have prototypes.
 +
 +Note that most of GCC is compiled such that warnings are errors,
 +so don't be sloppy!

codingconventions.html be an acceptable compromise?

Gerald